University of Groningen
Environmental and Financial Performance of Fossil Fuel Firms
Gonenc, Halit; Scholtens, Bert
Published in:
Ecological Economics DOI:
10.1016/j.ecolecon.2016.10.004
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.
Document Version
Final author's version (accepted by publisher, after peer review)
Publication date: 2017
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Gonenc, H., & Scholtens, B. (2017). Environmental and Financial Performance of Fossil Fuel Firms: A Closer Inspection of their Interaction. Ecological Economics, 132, 307-328.
https://doi.org/10.1016/j.ecolecon.2016.10.004
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
0
Environmental and Financial Performance
of Fossil Fuel Firms: A Closer Inspection of their Interaction
Abstract
We investigate the relationship between environmental and financial performance of fossil fuel firms. To this extent, we analyze a large international sample of firms in chemicals, oil, gas, and coal with respect to several environmental indicators in relation to financial performance for the period 2002– 2013. We find that these firms have significantly higher scores on environmental performance efforts than other firms. We use a simultaneous equations system to identify the direction of the relationship between environmental and financial performance of the firms. We find that environmental
outperformance has no impact on financial performance for chemical firms, reduces returns and risks for coal companies, has a mixed impact on returns in oil and gas, and reduces financial risks for oil and gas firms. Financial outperformance reduces environmental performance in all fossil fuel
(sub)industries investigated. Our findings mainly support the opportunistic view regarding the impact of financial returns, which holds that financial performance negatively impacts social performance. Regarding financial risk, we find support for the stakeholder perspective where good environmental performance is beneficial from a finance perspective. We conclude to substantial differences in the environmental-financial performance relationship along fossil fuel firms in different subindustries.
Keywords: Environmental performance; Financial performance; Fossil fuel firms; Corporate responsibility; Firm-level analysis.
1
1. Introduction
Given the impact of fossil fuels on climate change, it seems very relevant to investigate how the environmental performance of fossil fuel firms (firms in oil and gas, coal, and chemicals) relates to their financial performance. More specifically, is good financial performance associated with sound environmental performance, or is there a trade-off? Further, is this relationship the same along
different performance measures and (sub)industries? Answering these questions is important to assess the potential for changes in operations by fossil fuel firms to transform the energy system. Several studies find that energy-intense companies are punished by the stock market for poor environmental performance (see Patten, 1992; Kolk et al., 2001; Kollias et al., 2012). These studies usually focus on the impact of events on company reputation (see, e.g., Spence, 2011), but not on company operations and related cash flows. Scholtens (2008) and Lioui and Sharma (2012) investigate the potential reasons why there would be a link between environmental and financial performance. The former study finds that it is highly dependent on the way in which these performances are being measured. The latter finds a negative direct impact of environmental on financial performance but a positive indirect impact.
Our study specifically investigates environmental and financial performance of fossil fuel firms. As such, it tries to focus on a much more homogeneous category than understood by the concept ‘social performance’ and its equivalents, which also relates to governance, ethical, and social issues with firms. To be precise, we investigate environmental and financial performance in three subindustries: chemicals, coal, and oil and gas. We rely on both qualitative and quantitative environmental
performance indicators that are much more fine-grained than those used in the literature thus far. Further, we rely on different financial performance measures to avoid biases and to account for the underlying value structure of firms. We also address endogeneity and try to detect structural relations between environmental and financial performance. We find that fossil fuel firms have significantly higher scores for their environmental performance efforts relative to firms in other industries, but it shows that this is highly sensitive to (sub)industry classification. It will not come as a surprise that we also find that fossil fuel firms produce more waste and emissions than firms in other industries. Further, we find that environmental outperformance does have no impact on financial performance for
2
chemical firms, reduces returns and risks for coal companies, and has a mixed impact on returns in oil and gas, and reduces financial risks for firms in oil and gas. Financial outperformance reduces
environmental performance in all the types of fossil fuel firms investigated. This shows that there are substantial differences in the relationships studied for the different subindustries. These findings suggest that any policy approach should account for the value chain at the subindustry level, since a ‘one size fits all’ policy is likely to have very distorting effects and, hence, to be ineffective.
The remainder of this paper proceeds as follows. We first discuss the background of the relationship between financial and environmental performance of the fossil fuel firms (i.e. firms in oil and gas, coal, chemicals). Then, we introduce the data and methods employed in our analysis. Next, we report the results from the univariate analysis and show the estimation results of the regression models. Finally, we discuss our conclusions.
2. Background and Hypotheses
Bénabou and Tirole (2006, 2010) argue that there are basically three reasons as to why firms and institutions would want to behave in a responsible manner (please note that these responsibilities pertain to environmental, ethical, social and governance characteristics). The first is altruism, that is, ‘doing the right thing’. Here, the firm does incur costs to avoid or reduce externalities, but does not necessarily get something in return, such as lower expenses or higher revenues. The second reason is greenwashing, where the firm claims to behave in a responsible manner to gain benefits, but does not actually change the way it operates nor internalize externalities. The third reason is strategic behavior. Here, the firm makes an effort and incurs real costs to reduce externalities. However, it also succeeds in increasing its revenues from behaving in a responsible manner. Firms act on the basis of all three reasons, but may place different weightings on each of them, resulting in differing outcomes regarding social responsibility.
Views regarding the social (in a broad encompassing sense) responsibilities of companies mainly hold that their responsibilities go beyond maximizing shareholder returns, including a focus on the
environment, ethical conduct of business operations, and responsibility to stakeholders (Campbell, 2007). From this perspective, companies should adopt policies and practices that align with the wider
3
societal good (Matten and Moon, 2008). This approach aims at stakeholders like employees,
customers, suppliers, communities, regulators, and the environment. The social policies and practices of firms reflect a behavioral standard regarding their social responsibilities (Campbell, 2007). It appears that the results of company policies and practices may vary widely and bear no
straightforward relationship with financial performance (Dam and Scholtens, 2015). Furthermore, policies and practices regarding corporate responsibility often are not clearly defined and go beyond what is written into laws and regulations (Heal, 2008; Chatterji et al., 2009).
Two meta-studies that investigate the literature on the financial and responsibility performance of firms are Wu (2006) and Margolis et al. (2009). Wu (2006) researches the relationships between the financial and responsibility performance of firms (the latter relates to the environmental, social and governance performance of firms in general within the context of his research). This author arrives at several results: (1) there is a positive relationship between responsibility and financial performance indicators; (2) market-based measures are weaker predictors of responsibility than other financial measures, such as accounting indicators; and (3) perception-based measures report a stronger responsibility–financial performance relationship than performance-based measures. Margolis et al. (2009) find a small but statistically significant positive correlation between financial and social performance. One problem with such meta-analyses is that a lot of information gets lost and that studies are equally weighted despite huge differences in research design and quality.
Apart from methodological problems, indicators of social responsibility as well as those of financial performance widely differ among the studies included. Margolis et al. (2009) and Schulze and Trommer (2012) specifically mention this problem and the challenge of defining the responsibility construct. Indicators and measures of responsibility tend to capture either a single specific dimension, such as philanthropic donations or pollution control, or are broad appraisals of responsibility as a whole, like ratings. The issue of multi-dimensionality also plays a role with financial indicators (see Dam and Scholtens, 2015). For example, Gregory et al. (2014) mention that accounting measures are backward looking, and their objectivity and informational value is questionable. Stock market measures, by contrast, are much more forward-looking, with expectations of future cash flows and timing of these flows as well as risk embedded within the stock price (Gregory et al., 2014).
4
Based on Preston and O’Bannon (1997), Scholtens (2008) provides a brief overview as to why there might be a particular causal relationship between financial and environmental or social performance. There can be a negative link as the latter involves costs and therefore weakens the firm’s competitive position, suggesting there is a trade-off between the two. As such, environmental and social issues may conflict with value maximizing behavior. In addition, managers may engage with social and environmental issues from an opportunistic perspective which may conflict with stakeholder and shareholder objectives. the managerial opportunism theory. This approach states that ‘when financial performance is strong, managers may attempt to cash in by reducing social expenditure in order to take advantage of the opportunity to increase their own short-term private gains’ (Allouche and Laroche, 2005). This is a form of agency costs. It also works the other way around: when financial performance weakens, managers might engage in social programs to offset or justify their disappointing results. The opportunism approach follows agency theory. Here, one believes a manager, when possible, has an incentive to put private gains first. When financial performance is strong, managerial opportunism expects less social performance. Thus, the opportunism approach assumes that financial performance precedes social performance. Please not that there can also be a positive association. For example, satisfying stakeholders’ non-financial interests may result in improving the firm’s financial performance due to increased loyalty. Firms do have a social impact and there is a demand from stakeholders for responsible conduct of the firm and in equilibrium the costs and benefits of servicing this demand would cancel out.
As to the direction of the causality, there is the financial resources-based view where financial means are essential in order to invest in responsible conduct and performance (the availability of funds, hereafter ‘resources’). According to Orlitzky et al. (2003), the resource perspective suggests that investments in social performance may help firms develop new competencies, resources, and
capabilities which are manifested in a firm’s culture, technology, structure, and human resources (see also Russo and Fouts, 1997). Orlitzky et al. (2003) argue that social performance may help build managerial competencies because preventive efforts necessitate significant employee involvement, organization-wide coordination, and a forward thinking managerial style. They conclude that social
5
performance can help management develop better scanning skills, processes, and information systems, which increase the organization’s preparedness for external changes, turbulence, and crises. The same type of causality does occur in the more classical view of production which does occur to the
detriment of social welfare (i.e. the classical externalities).
The causality can also run from environmental to financial performance. This is the case with stakeholder theory (which assumes a positive relationship) and the trade-off perspective (which assumes a negative relationship). Stakeholder theory suggests that social performance is positively associated with financial performance because it enhances the satisfaction of various stakeholders - and consequently the firm’s external reputation – and leads to better financial performance (Allouche and Laroche, 2005). According to Preston and O’Bannon (1997), there is a lead-lag relationship between social and financial performance; external reputation (favorable or unfavorable) develops first, then financial results (favorable or unfavorable) follow.’’ According to Orlitzky et al. (2003) managers can increase the efficiency of their organization’s adaption to external demands by
addressing and balancing the claims of multiple stakeholders. Donaldson and Preston (1995) state that the widely believed notion is that stakeholder management contributes to successful economic
performance, but they add that this is insufficient to stand alone as a basis for the stakeholder theory. They state that ‘’studies have tended to generate implications suggesting that adherence to stakeholder principles and practices achieves conventional corporate performance objectives as well or better than rival approaches’’ (Donaldson and Preston, 1995).
As to the trade-off view, Preston and O’Bannon (1997) argue that social performance is the
independent variable and that social accomplishments involve financial costs. Allouche and Laroche (2005) mention that ’because social accomplishments involve financial costs, social responsibility may siphon off capital and other resources from the firm, putting it at a disadvantage compared to other firms that are less socially active. Lioui and Sharma (2008) assess the impact of environmental performance on financial performance as measured by return on assets and Tobin’s Q. They find a negative relationship between the two. However, they also detect a positive indirect effect as
6
Sharma, 2008). Pätäri et al. (2014) investigate how qualitative assessments of social responsibility “strengths and concerns” relate to the financial performance of 14 energy companies. They use Granger causality tests and find that social responsibility concerns Granger-cause corporate profitability and market value, whereas social responsibility strengths Granger-cause only market value. Furthermore, financial performance does not Granger-cause corporate social responsibility (Pätäri et al., 2014). However, they don’t investigate environmental performance and refrain from investigating financial risk, and rely on accounting information only.
Stock market returns are widely used to analyze financial performance in relation to corporate social responsibility (see Margolis et al., 2009). But studies based on this indicator can produce misleading results because, in an efficient market, returns may be expected to reflect only (unexpected) changes in corporate social performance. This is problematic, as there is evidence to suggest that social
responsibility indicators may be sticky (Chatterji et al., 2009). If social responsibility levels remain unchanged or if the changes are relatively small, then a returns-based study can give the impression that corporate social performance does not affect financial performance. But even when returns-based studies find some financial impact from social responsibility, care needs to be taken regarding
interpretation of the results. For example, El Ghoul et al. (2011) find that firms with high social responsibility have lower cost of capital. Long-run returns to firms with high social responsibility may be lower for a given expected future cash flow because they are subject to less market risk. Then, if social responsibility does lower a firm’s cost of capital, focusing solely on returns to indicate its financial impact will be misleading (Dam and Scholtens, 2015).
Understanding the overall financial implications of social responsibility requires that attention be given to both stock returns and firm value. To this extent, Dam and Scholtens (2015) provide underpinnings for the actual behavior of market participants. They relate social performance to measures like the market-to-book ratio (firm market value in relation to accounting value), return on assets, and stock market return. They conclude that there is a strong theoretical foundation for a positive relationship between social responsibility and financial performance, and argue that the relation is highly conditional on which financial performance measure is considered (Dam and
7
Scholtens, 2015). Gregory et al. (2014) argue that markets positively value most aspects of social responsibility, and do so because, in the long run and measured across most dimensions, high social responsibility firms have a higher expected growth rate in their abnormal earnings. But this seems to be due primarily to industry effects rather than to a particular social responsibility strategy. Therefore, it is important to investigate different financial performance measures alongside a host of
environmental indicators, and to focus on specific industries.
Heal (2008) argues that when a firm’s private and social costs are about the same, markets generally are beneficial for society. However, when corporate and social costs are not closely aligned, markets do not work so well for society. In this respect, the conflicts between corporations and society over environmental issues almost always derive from the external costs associated with pollution (Heal, 2008). Firms may try to internalize some of these external costs and, as such, act in a more socially responsible manner. In part, this results from pressure of the market and society and this is stronger when the firm operates closer to both of these (e.g., there will be more scrutiny on firms in the downstream of the supply chain than in the upstream). Further, it appears that, in relating
environmental performance to financial performance, it is important to pay attention to various types of indicator, as environmental performance is not a one-dimensional construct (Chatterji et al., 2009; Schulze and Trommer, 2012). Heal (2008) regards companies particularly in the tobacco, alcohol, pharmaceuticals, chemicals, and energy industries as facing great discrepancies between private and social costs. This is illustrated by Hong and Kacperczyk (2009), who find that investment portfolios consisting of firms in the tobacco, alcohol, and gambling industries in the US outperform portfolios without these industries. This suggests that these firms face higher cost of capital and incur more risk to attract investors.
We focus on the fossil fuel-intense firms (especially firms in oil and gas, coal, chemicals) and their environmental performance. Energy is a critical input to economic and societal processes and a part of all production processes. Thus far, several studies investigate the societal impact of energy companies. In this respect, they usually investigate disasters such as explosions or oil spills (e.g., Patten and Nance, 1998; Capelle-Blancard and Laguna, 2009). Further, the nature of these firms’ operations
8
requires high environmental exposure. Therefore, they are subject to economic, financial and political risks that are at least different in other industries. Thus, environmental outperformance can be
regarded as a strategy of energy firms to limit their exposure and/or to improve their reputation (Heal, 2008). Kolk and Levy (2001) show that energy firms invest resources in low-emission and renewable sources as well as in anticipating regulation to hedge themselves against exposure to the
environmental and societal impact of their operations.
From this broad overview of the literature, we arrive at several hypotheses we want to put to the test: First, based on the views of a.o. Kolk et al. (2001), Heal (2008), and Kollias et al. (2012), is that we want to find out whether environmental performance of our sample of fossil firms differs from that of other firms. Here, based on the literature discussed above, we hypothesize that their policies will be more intense and that they score relatively high on environmental policies (H1).
Second is that their actual performance in terms of emissions may be worse as this basically is the reason as to why they would engage more with environmental responsibility (H2). This would be reflected in much more efforts regarding emission reduction, product innovation and resource reduction of the fossil fuel firms.
Next, we assume that within this group of firms, the performance of chemical firms is superior to that of oil & gas and coal companies (H3). This is because chemical firms operate closer to the market of end-users and are more competitive than the energy industry (Budde, 2011). In this respect, Heal (2008) argues that firms that are more subject to the scrutiny of market participants are more likely to invest in responsibility. However, he relates this argument to broad-based industry classifications. Kolk et al. (2001) investigate reporting practices at the industry level and their study tends to confirm Heal’s view. We want to find out whether this also is the case for an industry that already is regarded as problematic. We don’t expect a significant difference between oil and gas companies vis-à-vis coal companies as they are more or less in the same position in this respect.
As to the relationship between financial and environmental performance, i.e. both the direction and the positive or negative relationship, we want to find out which of the different theoretical approaches in this respect would appear to hold (see Preston and O’Bannon, 1997; Scholtens, 2008).
9
As such, we want to test whether the stakeholder theory (H4), the trade-off view (H5), the resources approach (H6), or the opportunism view (H7) does hold for our samples.
Thus, in H4 we test stakeholder theory which assumes there is a positive impact from environmental to financial performance.
In H5 we test the trade-off hypothesis which holds that there is a negative impact from environmental to financial performance.
In H6 we test the resources approach which implies that financial performance is having a positive impact on environmental performance.
In H7 about the opportunism hypothesis which holds that financial performance will negatively impact environmental performance.
Further, in line with Pätäri et al. (2014), we try to find out what determines firms’ environmental performance and whether this differs for fossil fuel firms compared to other (i.e., non-fossil) firms.
3. Data and Method
We investigate environmental and financial performance of a large international sample of firms in both fossil fuel-related (firms in oil and gas, coal, chemicals) and ´non-fossil fuel-related´ industries (of course, we are well aware of the indirect usage of fossil fuels in all firms and in fact there is no industry that does not indirectly consume any fossil fuel) for the period 2002–2013. This period is motivated primarily on the basis of data availability of both the financial and the environmental variables. As to the fossil fuel firms, we include all firms in the following 2-digit SIC codes: 12 (‘coal’), 13 (‘oil and gas’), and 28 and 29 (‘chemicals’).
The quality of the ways in which responsibility is measured is a concern in the academic literature (see Chatterji et al., 2009; Schultze and Trommer, 2012). Most research on corporate social responsibility tends to rely on qualitative assessments from specialized ratings agencies. However, such assessment is usually based on specialist views regarding corporate policies and not so much on actual firm performance (Chatterji et al., 2009). Further, the assessment is not verified and cannot be replicated by outsiders. Since the relationship between policy and performance is not one-on-one, it would be better to use both types of indicators, namely, categorical assessment data and environmental performance
10
data for one specific industry, and to compare across industries. Therefore, we will want to use a wide array of indicators.
Our data about environmental performance are gathered from the Worldscope database provided by Thomson Reuters. The selection of sample firms is based on data availability at the ASSET4 ESG database in Worldscope. The literature is divided in terms of selecting data source to measure
environmental as well as social and governance scores when they use an international sample. Arouri and Pijourlet (2016) use Intangible Value Assessment ratings from MSCI and list the following studies to use the same: Derwall et al., 2005; Aktas et al., 2011; Guenster et al., 2011; Marsat and Williams, 2013. However, the coverage of ASSET4 ESG database has increased importantly, and therefore the choice of very recent studies (i.e., Cheng et al., 2014; Stellner et al., 2015; Feng et al., 2015; El Ghoul et al., 2016). We also feel ASSET4 is to be preferred due to the consistency in the reporting (e.g. MSCI is faced with a major structural break in the series in 2009). Further, the same provider, i.e. ThomsonReuters, also provides financial information about the companies. Therefore, it is likely that the matching errors will be much more limited than in the case of combining different data sources.
The ASSET4 ESG database carries historical data for several key performance indicators on four pillars: economy, environment, social, and corporate governance. The ASSET4 ESG framework allows us to rate and compare companies against approximately 700 individual data points, which are combined into over 250 key performance indicators. The scores on the key performance indicators are aggregated into a framework of 18 categories grouped within the four pillars, which are integrated into a single overall score. This database has gathered data from publicly available information, such as company websites, annual reports, and proxy files since 2002. Therefore, our analysis will cover the period 2002 to 2013. The coverage of the database originally was limited to US and European firms, but expanded in more recent years. As such, we have an unbalanced panel. We will report the results of our analysis for the overall sample in the main text, but we will also report them for subsamples of countries in the appendix and discuss these in the main analysis.
In our analysis, first we use overall percentage scores of the environment pillar (the Environmental Score), and extend our analysis to the three constituting categories of environmental performance: emissions reduction, product innovation, and resource reduction. Environmental score in fact measures
11
a company's impact on living and non-living natural systems, including air, land, and water, as well as complete ecosystems. It reflects how well a company uses best management practices to avoid environmental risks and to capitalize on environmental opportunities to generate long-term
shareholder value. The three constituting categories (emissions reduction, product innovation, resource reduction) are labeled categorical indicators in the remainder of this paper. Next, we employ variables that are much closer to actual environmental performance, such as emissions and expenses. Appendix A provides definitions of the environmental variables used.
Financial data also are collected from Datastream/Worldscope for firms with available environmental performance indicators. We eliminate financial firms to avoid issues of regulatory influence on these firms. We compute five financial performance variables from the same database; three relate to value and return and two to risk. As to the former, we investigate two market performance variables, namely stock market excess returns, the difference in the percentage change in the US dollar stock return between the beginning and end of a year and the annual local market index return, and Tobin’s Q, the ratio of (book value of total assets + market value of common equity − book value of common equity) to the book value of total assets. The accounting performance measure is the widely used return on equity, the ratio of net income to common equity. Further, and novel in this strand of the literature, we include two specific risk measures. The first is business risk, which measures firm earnings volatility as an unsystematic risk and is computed as the standard deviation of operating income ratio over three-year overlapping periods of the sample period. (Operating income ratio is the ratio of operating income, which is the difference between sales and operating expenses, to sales.) The second is Beta, which measures the firm’s systematic risk and is calculated using daily stock returns in each year by running regressions for the firm’s stock returns against local market index returns for each firm.
Worldscope data may contain errors, and thus all financial variables are winsorized at 0.01 and 0.99 to avoid outliers affecting results. Compared to previous studies (Patten and Nance, 1998; Capelle-Blancard and Laguna, 2009; Henriques and Sadorsky, 2010; Pätäri et al., 2014; Arslan-Ayaydin and Thewissen, 2015), our sample is highly international (it encompasses firms from over 50 countries), focuses on a more recent period, and uses a much wider scope of both financial and environmental indicators. More specifically, we include excess stock returns among the financial performance
12
measures, account for risk characteristics, and concentrate on both qualitative and quantitative aspects of environmental performance. As such, we feel we are able to arrive at a much more detailed
inspection of the interaction between environmental and financial performance.
We first perform univariate tests for comparisons of the means and the medians of the variables between the fossil fuel firms (chemicals, coal, oil and gas) and other firms. Standard t-tests for mean and non-parametric tests for median are used for statistical comparisons. Next, we concentrate on pooled ordinary least squares (OLS) regression estimations for the effects of financial performance variables on environmental performance scores. We investigate all industries within the economy (except for banks and other financial services providers) and control for the fossil fuel firms via a dummy variable. In this respect, we focus on firms’ overall environmental score and its constituent categories (emissions reduction, product innovation, and resource reduction). We do this for the specific environmental performance indicators as well. We use the interaction variables between the fossil energy industry dummy and financial performance variables to test whether the effects of financial performance variables on environmental performance are statistically different between the fossil fuel firms and those in other industries. In line with the literature, we control for size, which is the natural logarithm of book value of assets in US dollars, research and development expenditures (R&D) scaled by book value of total assets, financial leverage, the ratio of the total of short- and long-term debt to book value of total assets, and net working capital, the ratio of the difference between current assets and current liabilities to book value of total assets, to control the liquidity of firms. We use country and year fixed effects in all regressions.
In our research framework, we propose that financial performance determines environmental
performance, but we acknowledge that it could also be plausible, as documented in the literature, that environmental performance affects financial performance (see Margolis et al., 2009). In this case, the environmental performance equation contains an endogenous variable, financial performance, and vice versa. To address this reverse causality problem, as well as the possibility that some of the
independent variables are jointly determined, we create a system of structural equations, including two equations for environmental and financial performance, separately. To estimate the model, we perform a three-stage process for systems of simultaneous equations by using two-stage least squares (2SLS)
13
estimations for each equation. We produce the stage least squares (3SLS) estimates from a three-step process: In three-step one, we develop instrumental variable equations for both environmental and financial performance variables. The two instruments in the environmental performance equation are averages of the scores by country/year and by country/industry score (Cheng et al., 2014). We use leverage and net working capital as instruments of financial performance (see Vishani and Shah, 2007; Afza and Nazir, 2009). We use all other variables explained in the OLS regression analysis above as control variables along with country and year fixed effects, and expect them to impact the relationship in line with the literature (Wu et al., 2006; Margolis et al., 2009). These two equations create the predicted values resulting from a regression of each endogenous variable on all exogenous variables in the system; this is identical to the first step in conventional 2SLS. Thus, the 3SLS process creates a consistent estimate for the covariance matrix of the equation disturbances. These estimates are based on the residuals from the 2SLS estimation of each structural equation. In the last step of the third stage, the 3SLS performs a generalized least squares (GLS) type estimation using the covariance matrix estimated in the second stage, and with the instrumented values in place of the right-hand-side endogenous variables.
4. Results
We first present the descriptive statistics and the univariate analysis. Then, we provide the findings from the regression analyses.
4.1 Univariate analysis
Table 1 is an overview of the country composition of the sample. It shows that in the 51 countries under investigation, there are more than 23,000 firm-year observations, among which about 12% are fossil fuel firms. Most observations are for the US, Japan, and the UK; together the three make up 53% of total observations (that is also the main reason why will provide estimation results for subsamples in the Appendix, namely for Australia, Canada, Japan, UK, and US, for the sample
excluding the UK and the US, and for the full sample excluding the US). Table 1 reports the means for the overall environmental score. Please see Appendix A for the definition of all the variables used in this construct. The environmental score is a performance pillar reflecting how well—according to the
14
rating agency—a company uses best management practices to avoid environmental risks and to capitalize on environmental opportunities to generate long-term shareholder value. A higher score relates to relatively more (perceived) efforts by the firm. However, this score does not necessarily imply that the firm is cleaner or better from an environmental perspective. To that extent, one has to investigate the actual performance indicators, which will be done later on in this study.
Table 1 reveals that the environmental score is higher for the 2,739 fossil fuel firms (comprising firms in oil and gas, coal and chemicals) compared to the 20,568 non-fossil fuel firms (54.1 versus 51.6). This suggests support for the first hypothesis where it was assumed that fossil fuel firms would outperform others in this respect. Table 1 shows that fossil fuel firms in Austria, Belgium, Finland, Hungary, and Italy have the highest environmental scores, whereas those in Ireland, Singapore, and Sweden have the lowest.
[Insert Table 1 about here]
Table 2 sets forth the sample composition for the 44 industries and their performance with respect to the main variables of interest. This table shows that most observations are for firms in oil and gas, business services, and in retail. The sectoral distribution of the observations is much less skewed than in the case of the country distribution: the three largest (oil and gas, retail, and business services) make up 20% of the total sample. Table 2 shows that the mean of the encompassing environmental score is relatively high (above 66) in aircraft, automobiles, computer hardware, business supplies, electronic equipment, consumer goods, chemicals, and recreation. It is relatively low (33 or less) in agriculture, defense, entertainment, personal care, precious metals, healthcare, and other industries. Among the fossil fuel firms, there is a marked difference between chemicals (68.7) on the one hand, and coal (39.1) and oil and gas (46.8) on the other. This is confirmation for the third hypothesis about the relative performance of fossil fuel subindustries. We want to point out that these findings align only to some degree with the general view put forward by Heal (2008). Industries with substantial
externalities, such as the aircraft, auto, chemical, machinery, rubber, shipping, steel, and tobacco industries, indeed score relatively high on the environmental score. However, this also is the case with industries where the differential between social and private costs seems much less obvious, including the computer hardware, business supplies, and recreation industries.
15
As to financial performance, Table 2 shows that the Tobin’s Q of chemicals, and oil and gas is relatively low. Their excess stock return is higher than that of coal firms and of firms in most other sectors/industries. For coal, the excess stock return is below the average of non-fossil fuel industries. Return on equity is about the same in the three fossil fuel-related sectors and slightly lower than with non-fossil fuel firms. The financial risk indicators reveal that most indicators for all three types of fossil fuel firms sectors are much higher than those elsewhere. The exception is business risk in chemicals, which appears low compared to the average of the non-fossil fuel firms. In general, these findings are in line with those found elsewhere (e.g., Schultze and Trommer, 2012; Pätäri et al., 2014; Arslan-Ayaydin and Thewissen, 2015).
[Insert Table 2 about here]
More detailed descriptives are shown in the six panels in Table 3, which also reports the median performances and provides more information regarding firm characteristics and environmental indicators. Furthermore, this table reports the test results regarding the differences between the mean and median performance of different subgroups (i.e., fossil intense firms and non-fossil fuel-intense firms; chemicals versus coal and oil and gas). Panel A in Table 3 compares the main financial characteristics. It shows that fossil fuel firms have lower Tobin’s Q, higher excess stock market returns, are more risky, are much larger, have less R&D as well as less working capital, and have slightly lower leverage. In most cases, the differences are statistically significant with 99% confidence, both in the means and medians (except leverage). Return on equity does not significantly differ
between fossil fuel firms and the other firms.
Panel B reports the differences between overall environmental score and the three other categorical indicators (emission reduction, product innovation, resource reduction). In this respect, the fossil fuel firms perform significantly better on overall environmental score and on efforts toward emission reduction, but do not significantly differ from other firms with respect to product innovation and resource reduction. Therefore, regarding the environmental score in general and the emission reductions in particular, we find support for H1, but not for H2. We don’t find this in the case of product innovation and resource reduction.
16
Panel C gives details of the financial characteristics of firms in the three fossil fuel-related sectors and compares chemical firms with coal and oil and gas firms. This panel shows that median Tobin’s Q and excess stock market return is significantly higher with chemicals but that the mean is not. Further, there is no statistically significant difference regarding stock market returns, and only a marginally significant (10% significance) difference for the median of return on equity. However, risk in coal and in oil and gas is much higher than that with chemicals. R&D and working capital are lower with coal and oil and gas firms compared to chemical firms; also, the former (especially oil and gas) are much larger than chemical firms.
Panel D provides an overview of the univariate tests of the four categorical environmental indicators for the three sectors. This panel clearly shows that chemical firms have much better environmental performance scores than those in coal, oil, and gas. This is supportive for H3 regarding the
subindustries in fossil.
Panel E shows the performance of fossil fuel firms compared to other firms for a large number of environmental performance indicators. This panel shows that the fossil fuel firms exhibit greater use of resources, water, and energy, and generate more emissions of all types. This is clearly in support of our second hypothesis. The mean for their NOx and SOx emissions and their waste production is lower than in non- fossil fuel firms, but the median shows they are higher. The mean of the fossil fuel firms regarding the amount of waste is lower than with the non- fossil fuel firms, but the median does not confirm this. Resource-use reduction policies and monitoring in the non-fossil fuel firms are seen as superior to those with the fossil fuel firms. This contrasts with H1.
Panel F in Table 3 shows the performance on environmental indicators of different types of fossil fuel firms: It compares firms in chemicals with those in coal and in oil and gas. This panel shows that the latter have higher environmental expenditures and environmental provisions (in line with Heal, 2008). Coal and oil and gas firms also have higher NOx, SOx, and volatile organic compound (VOC)
emissions than chemical firms. For most other environmental indicators (e.g., CO2 equivalent
emissions, water use, waste production, and energy use), chemical firms put more pressure on the environment. But this sector’s emission reduction efforts rate better than those in the oil and gas and in coal. The policies, implementation, and monitoring of emission reduction of chemical firms is
17
perceived as better than that of the coal and the oil and gas firms. As such, these findings confirm H3 and are in line with those of Pätäri et al. (2014) and Arslan-Ayaydin and Thewissen (2015) on the basis of KLD MSCI data.
[Insert Table 3 about here]
4.2 Regression analysis
Table 4 reports the estimation results of the OLS regressions, where the environmental category proxies are regressed against financial variables. Thus, first, we try to explain what determines the overall environmental score and its three categorical components, namely, emissions reduction, product innovation, and resource reduction. The model accounts for a dummy variable to compare fossil fuel firms with non- fossil fuel firms, and interaction effects. As such, we investigate the impact of financial performance of a firm on environmental score, conditional on the firm belonging to one of the three fossil fuel-related sectors. Apart from the five financial performance indicators, we use firm size, leverage, R&D expenditure, and working capital as control variables, as in many studies on the relationship between financial and social performance (see Wu et al., 2006; Margolis et al., 2009). All regressions are run controlling for country and year fixed effects.
The estimated coefficients of the dummy variable representing fossil fuel firms (Dummy_Fossil) show that this variable is indeed a significant factor for the overall environmental score, and that it
specifically relates to the emission reduction categorical score and to the product innovation category, which supports both H1 and H2. Further, Table 4 reveals that there is a mixed picture regarding how the financial performance indicators and the control variables relate to the different environmental categories. Tobin’s Q is positively associated with environmental performance, but the significant coefficient of the interaction term reveals that the relationship is, in fact, a negative one for fossil fuel firms. This implies that firms that are relatively highly valued are associated with relatively low environmental categorical scores. This suggests that with fossil fuel firms there is a trade-off regarding firm value and environmental performance, which confirms the opportunism hypothesis (H7). Excess stock returns are negatively related to the overall environmental score and to the resources category. Here, we find that the interaction with energy is statistically significant and there is a positive
18
relationship between excess returns and environmental performance. This supports the resources hypothesis (H6). For our third measure, return on equity, we find that these returns have a positive impact on the environmental performance scores but if we interact with the fossil fuel dummy, there is no significance. In general, these findings are in line with the predictions of Gregory et al. (2014) and Dam and Scholtens (2015). For business risk, there is a negative and significant relationship with environmental performance but if we investigate the interaction with Dummy_Fossil, it shows that this doesn’t have a significant role to play. For Beta as a risk indicator, we find that there is a statistically significant and positive relationship with the environmental categories. But, as with business risk, we don’t find that fossil fuel as such has an impact here (apart from a marginal negative impact regarding the product category). As to the controls, size clearly and positively contributes to a high score on the categorical environmental indicators, as do R&D and availability of net working capital. However, again, leverage is not significantly associated with environmental performance.
[Insert Table 4 about here]
In Appendix B, we show the estimation results of the same model used to arrive at the findings in Table 4, but focus on geographic subsamples. Appendix B.1 gives the results for a sample of
Australia, Canada, Japan, the UK, and the US, who make up about two thirds of the total sample. This shows that the relationships are much weaker than in the overall sample. Here, there is only weak support in the case of Tobin’s Q and there is no longer a positive relationship between excess stock returns in energy and environmental performance. Therefore, we conclude that there is no longer support for the hypotheses. Further, it shows that business risk in energy positively impacts environmental product performance. Appendix B.2 shows the results when the US and the UK are excluded, which renders 60% of the total sample. Here, the results are very much in line with those for the overall sample as depicted in Table 4 and we again find strong confirmation for H7, but less so for H6. Another interesting difference is that for Beta as the risk indicator, it clearly shows that more risk reduces environmental performance. Appendix B.3 shows the results when we exclude the US, which leaves us with about 70% of the original sample. These results are basically in line with those of the previous sensitivity analysis.
19
Table 5 provides the estimation results of the pooled OLS regression for the different types of fossil fuel firms (chemicals, coal, oil and gas). It shows detailed information on how financial performance is associated with the environmental score. In general, the estimation results show that the relationship between environmental performance and financial performance differs among the three groups. The results suggest that being among the fossil fuel firms as such does not produce a straightforward impact on the overall environmental score. We find that only with oil and gas firms there is a significant and negative association between Tobin’s Q and environmental score. This confirms the opportunism hypothesis (H7) for oil and gas firms. This result also seems to suggest that the negative relationship detected in Table 4 between this variable and environmental score is due to firms in the oil and gas sector in particular. We establish a significant and positive relationship between excess stock returns for oil and gas firms, but not for chemical firms and coal firms. Hence, we can conclude that H6 is supported for oil and gas companies, but not for the others. Table 5 shows that for firms in the coal industry, there is a statistically significant negative relationship between return on equity and environmental score. This too hints at a trade-off between financial and environmental performance and confirms H7. As to business risk, there is a clear positive association between this risk indicator and environmental performance for coal firms as well as for firms in oil and gas, but not for chemical firms. This suggests that particularly the relatively risky firms have higher environmental scores. For Beta, we observe that belonging to the oil and gas sector implies a significant negative relationship between this financial market risk (business risk) and environmental score. There is no significant relationship between the Beta of a coal firm and this score, whereas there is a significant positive one between the Beta of a chemical firm and the environmental score. The controls, again, show a significant and positive association with the dependent variable, with leverage the exception.
[Insert Table 5 about here]
In Appendix C, we redo the estimations for Table 5 for three different subsamples. Appendix C.1 reports the results for Australia, Canada, Japan, UK and US. It shows that there is no significant association between Tobin’s Q, excess returns and return on equity interacted with energy for any of the three subindustries. Hence, there is no support for the resources or opportunism hypotheses. As to risk, we find that there is a significant positive relationship between business risk and environmental
20
score for the coal firms, and marginally so for oil and gas companies. There is a marginally negative relationship with chemical firms in this respect. But for Beta, there is a positive association between risk and environmental score with chemicals. Appendix C.2 with the results for the sample excluding the UK and the US, as well Appendix C.3 excluding US only, render highly similar results to those in the main analysis reported in Table 5.
In Table 6, we report the last phase of the estimation, namely, the 3SLS estimation of our
simultaneous equation system regarding the effect of fossil fuel on environmental performance, as well as that of environmental score on financial performance. This is motivated by the endogeneity of the relationship between environmental and financial performance, as is widely documented in the literature (see the reviews: Wu et al., 2006; Margolis et al., 2009). In all panels, we first have financial performance as the dependent variable and environmental performance as the independent variable in the upper half; this is reversed in the lower half of the panels.
Panel A in Table 6 shows the estimation results for all three fossil fuel-related industries combined, and panels B–D show those for chemical, coal, oil and gas firms, respectively. The overall results in Panel A suggest that environmental performance of fossil fuel firms
(Dummy_Fossil*EnvironmentalPerformance) has a positive impact on Tobin’s Q, but no significant impact on return on equity. This lends support for the stakeholder hypothesis (H4) in the case of Tobin’s Q only. Further, environmental performance in fossil fuel firms significantly reduces excess returns and both risk measures, which is in support of the stakeholder hypothesis (H4). When financial performance is the independent variable, this shows that there is a statistically significant (<1%) and negative relationship with all financial variables, except return on equity and Beta
(Dummy_Fossil*Financial Performance). This is understood as follows: the estimated coefficient of Dummy_Fossil is positive and mostly significant, suggesting that fossil fuel firms are to be associated with relatively higher environmental scores. Financial performance also yields a positive and
significant sign, except for ROE and Beta. This suggests that better financial performance is associated with better environmental performance. But the combination of the two yields a statistically significant and negative sign for again three proxies of financial performance. This implies that, for fossil fuel firms that perform relatively well from a financial perspective, there is a significant and negative
21
association with environmental score. This confirms the opportunism hypothesis (H7). For the two risk indicators, panel A in table 6 shows there is a negative and significant association with business risk but no relationship with market risk (i.e., Beta). This confirms the resources view (H6) for business risk only.
Panels B–D show the results for fossil fuel firms in the three groups, chemicals, coal, and oil and gas, respectively. In the discussion, we again focus on the interaction coefficients in the upper and lower half of the three panels. Panel B shows that with environmental performance as the independent variable, there is no statistical significance for the coefficient of the interaction term, suggesting that environmentally outperformance of chemical firms has no impact on financial performance for any of the five performance measures. With financial performance as the independent variable, the result is quite different. Financially outperforming chemical firms have a significant negative impact on Tobin’s Q, stock market returns, and return on equity. There is no significant relationship with the two financial risk measures. Again, this is supportive of the opportunism hypothesis (H7) which posits a negative relationship between financial and social performance.
Panel C shows the results for coal firms. These are quite similar to those in Panel B. However, there now is some marginal significance for the interaction coefficient when environmental performance is the independent variable regarding the Tobin’s Q and Beta. And with return on equity with coal firms, we find support for the trade-off hypothesis (H5) where social outperformance is negatively associated with financial performance. When we use financial performance as our independent, the firms in the coal industry show the same behavior as those in chemicals, and we confirm H7.
The results for firms in oil and gas (Panel D) are quite different from those for chemicals and coal. This panel shows that when environmental performance is the independent variable, there is a significant positive impact on Tobin’s Q (confirming the stakeholder view, H4), and a significant negative impact on excess stock market returns (confirming the trade-off view, H5) as well as on the two risk measures (again confirming the stakeholder view, H4). This suggests that environmentally outperforming firms in oil and gas have relatively higher value and lower stock market returns, as well as lower financial risk. Further, with financial performance as the independent variable and
22
that perform well from the financial perspective have significantly lower Tobin’s Q and excess stock market returns. This is strong support for the opportunism hypothesis (H7). There is no significant relationship with return on equity. In addition, these firms are associated with less business risk (confirming H6), but not with systematic risk. In all cases, this panel shows that the effects of our controls are highly statistically significant.
[Insert Table 6 about here]
Sensitivity analysis regarding the 3SLS estimations for the simultaneous equation system for the effect of the fossil fuel (sub)industries for geographic subsamples is reported in Appendix D. Appendix D.1 has the results for the subsample of Australia, Canada, Japan, the UK and the US. The overall results are similar to those in Panel A of Table 6, but the panels B-D all reveal that better environmental performance significantly reduces Tobin’s Q with chemicals and coal, but it improves Tobin’s Q with oil and gas companies. Regarding the risk measures, the results of this sensitivity analysis in general as in the same direction as in the main analysis but significance is weaker. Appendix D.2 shows the results for our sample excluding the US and the UK. These findings are highly similar to those in the main analysis as reported in Table 6.
We establish that the fossil fuel firms in general outperform other firms regarding the overall environmental score. This seems to be based on these firms’ efforts to behave in a more responsible manner. However, we also establish that this especially results from outperformance by chemical firms. Firms in coal and in oil and gas significantly underperform firms in most other industries. If we investigate the relationship between financial and environmental performance for these three fossil fuel related sectors, we find that industry specifics is mainly of importance in the risk arena. Further, selection of the dependent and independent variables does matter. This is especially the case with the value and return indicators. If environmental performance is used as the independent variable, there is a significant positive relationship between environmentally outperforming firms in oil and gas and Tobin’s Q, and a negative one for these firms and excess stock market returns. This confirms the predictions from the theoretical model of Dam and Scholtens (2015). We also discover a negative association between environmental outperformance and market risk which especially is the case with oil and gas firms. With financial performance as the independent variable, we observe a statistically
23
significant and negative association with all three financial indicators in all three subindustries (apart from return on equity in oil and gas). Subindustry specifics, however, again clearly show up in the risk arena.
5. Discussion and Conclusion
We study the performance of a large, international sample of companies that are highly intense regarding the use of fossil fuels with respect to several environmental dimensions of corporate social responsibility in the period 2002–2013. We relate their environmental performance to various measures of corporate financial performance. The fossil fuel firms are of particular interest as their social costs are substantially above their private costs: External effects are a major concern with these firms. In particular, the role of fossil fuels in climate change is a topic of intense interest and debate. This is one of the main reasons fossil fuel firms place great effort into improving their social and environmental policies and performance (Kolk et al., 2001; Heal, 2008). We investigate how environmental performance relates to fossil fuel firms’ financial performance. As to environmental performance, we use qualitative and quantitative information from Thomson Reuter’s ASSET4. For financial performance, we investigate different, mostly hitherto unexplored, financial indicators relating to stock market and accounting performance, namely, Tobin’s Q, excess stock returns, return on equity, business risk, and Beta (systematic risk). It shows that there is a lot of heterogeneity within our sample, both regarding indicators of environmental and financial performance.
We find that, in most instances, there is a strong and significant relationship between corporate environmental and financial performance of the fossil fuel firms. This especially concerns Tobin’s Q and return on equity. For excess stock market returns, there usually is no relationship or only a small negative effect. In general, we conclude that when firms do well regarding Tobin’s Q and return on equity, they also show high environmental scores. It should be remembered that these scores pertain to policies to a great extent. When we account for the fact that a firm operates in a particular fossil fuel-related sector (chemicals, coal, or oil and gas), this characteristic plays a very crucial role. Operating in the fossil fuel-related industry as such appears to change the general relationship between
environmental and financial performance: there no longer is a significant and positive relationship and, especially in the case of Tobin’s Q, there appears to be a statistically significant and negative
24
association. When we investigate a wide range of environmental performance indicators (e.g., emissions, water use, waste production, resource use) it shows that better financial performance implies more emissions. This suggests that there is a trade-off between Tobin’s Q and environmental performance with fossil fuel firms, reflecting the external effects of their production processes. In general, we infer that particularly financial outperformance matters. Environmental outperformance is not associated with Tobin’s Q, stock market returns, or return on equity with firms in chemicals and coal. However, with oil and gas firms, we find that environmental outperformance is significantly associated with Tobin’s Q and stock market returns. Furthermore, in oil and gas, both environmental and financial outperformance can be associated with lower risk.
We tested several hypothesis. We found support for the notion that fossil fuel firms have better policies (H1) but weaker actual performance (H2) than non-fossil ones. However, H2 is to be rejected for the chemical industry. We can also confirm our H3 which holds that environmental performance of firms in chemicals is better than that of firms in coal and in oil and gas, but only in a univariate setting. As to the relationship between financial and environmental performance, we found some support for the stakeholder theory (H4), especially in the risk dimension, which is well in line with the findings of Scholtens (2008). Furthermore, we find little support for the trade-off view (H5), which holds that social performance goes to the detriment of financial performance, apart from firms in coal. The same is the case with the resources hypothesis (H6), which assumes that financial performance has a positive impact on social performance. This especially seems to be the case for companies in oil and gas, where it shows that social performance significantly reduces the risk indicators (business risk and Beta). However, in most cases, we found that financial performance has a significant negative impact on social performance, as such confirming the opportunism hypothesis (H7).
Our findings contribute to the literature in several ways. First, we convincingly show that risk
management is an issue in oil and gas; this is in line with the theoretical notions of Bénabou and Tirole (2006, 2010). Secondly, we find that industry-specific issues are important, as discussed by Heal (2008) and are able to show in more detail how they are so. Further, we add to the literature on the relationship between finance and corporate social responsibility, as discussed by Wu et al. (2006) and Margolis et al. (2009), for a much broader range of environmental and financial indicators than has
25
been studied so far for firms in chemicals, coal, and oil and gas. Fourth, we illustrate the notions brought forward in the more theoretical studies of Gregory et al. (2014) and Dam and Scholtens (2015) via a case study of fossil fuel firms. Our study also complements the findings about the relationship between financial and environmental performance with energy firms, which thus far investigate mainly accidents (Patten and Nance, 1998; Capelle-Blancard and Laguna, 2009; Henriques and Sadorsky, 2010). Another novelty is that we show that the inclusion of financial risk sheds light on the relationship between environmental and financial performance, thus complementing Pätäri et al. (2014) and Arslan-Ayaydin and Thewissen (2015). Finally, we develop a broad international setting and perspective, as our data include more than 50 countries.
From a policy perspective, our findings suggest that environmental policies for fossil fuel firms need to account for industry-specifics, as a one size fits all approach is unlikely to achieve policy objectives. Policy design should be very careful as to what specific objective is targeted, given the complex relationship between environmental and financial performance in the fossil fuel-related sectors.
Limitations of our study include the quality of the environmental performance data, as these are not externally verified and validated. We regard this as an important drawback regarding scientific research in this area and very much welcome initiatives to overcome this problem. Further, there is a bias in our study toward observations from richer countries. Although we include many more observations from developing countries than is the case in previous studies, we would like to investigate whether the relationships differ among subgroups of countries as well. Our sensitivity analyses show that there in some instances sample composition has an impact on the conclusions. Regarding the methodology, we rely on a 3SLS approach that is subject to some weaknesses. In particular, finding the best instruments that impact environmental (financial) performance, but not financial (environmental) performance, is very difficult. Even though there may be validity arguments against our model and instruments, they are consistent with those used in the literature.
We conclude that efforts of fossil fuel firms do not appear sufficient to improve environmental performance and that there are both trade-offs and synergies between environmental and financial performance, which can differ along the various indicators and which are highly industry-specific.
26
References
Afza, T., & Nazir, M., 2009. Impact of aggressive working capital management policy on firms’ profitability. The IUP Journal of Applied Finance, 15, 19-27.
Aktas, N., de Bodt, E., & Cousin, J.-G., 2011. Do financial markets care about SRI? Evidence from mergers and acquisitions. Journal of Banking & Finance, 35, 1753–1761.
Allouche J., & Laroche, P., 2005. A meta-analytical investigation of the relationship between
corporate social and financial performance. Revue de Gestion des Resources Humaines, 57, 18– 41.
Arouri, M., & Pijourlet, G.J., 2016. CSR performance and the value of cash holdings: International evidence. Journal of Business Ethics, doi:10.1007/s10551-015-2658-5
Arslan-Ayaydin, Ö., & Thewissen, J., 2015. The impact of environmental strengths and concerns on the accounting performance of firms in the energy sector, in Dorsman, A., Westerman, W., & Simpson, J.L. (eds.), Energy technology and valuation issues, New York, Springer, 83-107. Bénabou, R., & Tirole, J., 2006. Incentives and prosocial behavior. American Economic Review, 96,
1652-1678.
Bénabou, R., & Tirole, J., 2010. Individual and corporate social responsibility, Economica, 77, 1-19. Budde, F., 2011. Chemicals’ changing competitive landscape. McKinsey & Company,
http://www.mckinsey.com/industries/chemicals/our-insights/chemicals-changing-competitive-landscape.
Campbell, J.L., 2007. Why would corporations behave in socially responsible ways? An institutional theory of corporate social responsibility. Academy of Management Review, 32, 946-967. Capelle-Blancard, G., & Laguna, M.A., 2009. How does the stock market respond to chemical
disasters? Journal of Environmental Economics and Management, 59, 192-205.
Chatterji, A.K., Levine, D.I., & Toffel, M.W., 2009. How well do social ratings actually measure corporate social responsibility? Journal of Economics and Management Strategy, 18, 125-169. Cheng, B., Ioannou, I., & Serafeim, G., 2014. Corporate social responsibility and access to finance.
Strategic Management Journal, 35, 1-23.
Dam, L., & Scholtens, B., 2015. Towards a theory of responsible investing: On the economic foundations of corporate social responsibility. Resource and Energy Economics, 41, 103-121. Derwall, J., Guenster, N., Bauer, R., & Koedijk, K., 2005. The eco-efficiency premium puzzle.
27
Donaldson T., & Preston, L.E., 1995. The stakeholder theory of the corporation: concepts, evidence, and implications. Academy of Management Review, 20, 65–91
El Ghoul, S., Guedhami, O., Kwok, C.C.Y., & Mishra, D.R., 2011. Does corporate social responsibility affect the cost of capital? Journal of Banking and Finance, 35, 2388-2406. El Ghoul, S., Guedhami, O., & Kim,Y., 2016. Country-level institutions, firm value, and the role of
corporate social responsibility initiatives. Journal of International Business Studies, doi:10.1057/jibs.2016.4.
Feng, Z.-Y., Wang, M.-L., & Huang, H.-W., 2015. Equity financing and social responsibility: Further international evidence. International Journal of Accounting, 50, 247-280.
Gregory, A., Tharyan, R., & Whittaker, J., 2014. Corporate social responsibility and firm value: disaggregating the effects on cash flow, risk and growth. Journal of Business Ethics, 124, 633-657.
Guenster, N., Bauer, R., Derwall, J., & Koedijk, K., 2011. The economic value of corporate eco-efficiency. European Financial Management, 17, 679–704.
Heal, G., 2008. When principles pay. Corporate social responsibility and the bottom line. New York and Chichester, Columbia Business School Publishing.
Henriques, I., & Sadorsky, P., 2010. Can environmental sustainability be used to manage energy price risk? Energy Economics, 32, 1131-1138.
Hong, H., & Kacperczyk, M., 2009. The price of sin: The effects of social norms on markets. Journal of Financial Economics, 93, 15–36.
Kolk, A., & Levy, D., 2001. Winds of change: Corporate strategy, climate change and oil multinationals. European Management Journal, 19, 501-509.
Kolk, A., Walhain, S., & Van de Wateringen, S., 2001. Environmental reporting by the Fortune Global 250: Exploring the influence of nationality and sector. Business Strategy and the Environment, 10, 15-28.
Kollias, C., Papadamou, S., & Stagiannis, A., 2012. The financial spillovers of the Gulf of Mexico oil accident. Working paper, University of Thessaly.
Lioui, A., & Sharma, Z., 2012. Environmental corporate social responsibility and financial performance: Disentangling direct and indirect effects. Ecological Economics, 78, 100-111. Margolis, J.D., Elfenbein, H.A., & Walsh, J.P., 2009. Does it pay to be good … and does it matter? A
meta-analysis of the relationship between corporate social and financial performance. Available at SSRN: http://ssrn.com/abstract=1866371.
